Bruce Bradbury | Lightnews

John Holbein @johnholbein1.bsky.social · 7d

This is terrifying.

"[AI agents] can... infer a researcher's latent hypotheses and produce data that artificially confirms them."

...

"We can no longer trust that survey responses are coming from real people" [email protected]

Bruce Bradbury @brucebradbury.bsky.social · 8d

How to not sound like a LLM - and write something interesting.

Ryan Cordell @ryancordell.org · 9d

By far one of the most impactful readings of the semester in "Writing with Robots" has been @johnrgallagher.bsky.social's piece on LLM’s propensity to substitute lists for argumentation—students are suddenly *noticing* the lists everywhere & engaging them critically

The Curious Question of AI-written Lists: Or, LLMs are Genre Machines

Ending with a solution for teaching writing

meresophistry.substack.com

Ryan Cordell @ryancordell.org · 9d

By far one of the most impactful readings of the semester in "Writing with Robots" has been @johnrgallagher.bsky.social's piece on LLM’s propensity to substitute lists for argumentation—students are suddenly *noticing* the lists everywhere & engaging them critically

The Curious Question of AI-written Lists: Or, LLMs are Genre Machines

Ending with a solution for teaching writing

meresophistry.substack.com

NBER @nber.org · 10d

Exposure to the Low Income Housing Tax Credit during childhood improves mother's health and socioeconomic status at the time of first birth as well as leading to improved infant health outcomes, from Janet Currie and Jessica Van Parys www.nber.org/papers/w34464

Bruce Bradbury @brucebradbury.bsky.social · 14d

e61 have a report arguing for a broadening of the Australian GST tax base. However, looking at variation across expenditure levels is misleading because of 'shopping variability'. People with high expenditures tend to be those who have just purchased consumer durables.
e61.in/are-gst-exem...

Effective GST tax rates as a share of total spending. Ordered by spending and income percentiles.

Greg Jericho @grogsgamut.bsky.social · 15d

Gough raised unemployment benefits to their highest level relative to poverty - then Fraser (and his treasurer John Howard) fucked it up as soon as they could.

Hawke/Keating repaired some of the damage... then Howard fucked it up for good
thepoint.com.au/opinions/251...

Graph showing unemployment benefits relative to poverty

Bruce Bradbury @brucebradbury.bsky.social · 20d

Income support recipients still have the highest annual price rise (Australia, Sept qtr 2025).

John Quiggin @johnquiggin.bsky.social · 20d

This is the kind of outcome the National Electricity Market, combined with smart meters and retail competition, was supposed to deliver. Instead, the government has had to impose it. Yet another patch on a failed system #auspol

Australian households to get free electricity three hours a day

Saying there is enough solar power for everyone in the daytime, the federal government will direct retailers to provide three hours of free power every day to consumers.

www.abc.net.au

Bruce Bradbury @brucebradbury.bsky.social · 25d

Big growth in Australian average incomes up until 2009 (+small increase in inequality). Flat thereafter. Note, the latest observation for Australia would be 2017-18 or 2019-20, not 2024.

Bruce Bradbury @brucebradbury.bsky.social · Oct 26

Big mistake. Being an insightful discussant is the best (and possibly easiest) way to impress your fellow academics.

Bruce Bradbury @brucebradbury.bsky.social · Oct 20

"People are not rebelling against economic elites, but rather against cognitive elites... Seeing things in this way makes it easier to understand why people get so worked up over seemingly minor issues, like language policing."
josephheath.substack.com/p/populism-f...

Populism fast and slow

It is natural that a person who is both concerned by the rise of right-wing populism and possessed of a bookish disposition might turn to the academic political science literature in search of a bette...

josephheath.substack.com

Bruce Bradbury @brucebradbury.bsky.social · Oct 12

I'll be talking about Australian poverty trends tomorrow:

Bruce Bradbury @brucebradbury.bsky.social · Sep 30

This doesn't necessarily mean that older people are less happy. For the average person, they have less life satisfaction as they age. But the most unhappy ones die off - leaving the 'happy' ones behind.

Tamkinat Rauf @tsrauf.bsky.social · Sep 29

Life satisfaction mostly declines with age. Previous findings (esp. the famous U-shaped age-SWB trajectory) were artifacts of misspecified models. doi.org/10.1093/esr/...

Tamkinat Rauf @tsrauf.bsky.social · Sep 29

Life satisfaction mostly declines with age. Previous findings (esp. the famous U-shaped age-SWB trajectory) were artifacts of misspecified models. doi.org/10.1093/esr/...

Bruce Bradbury @brucebradbury.bsky.social · Sep 25

Unfinished business: 50 years after Henderson (14 October 2025). Online panel discussion on the trajectory of Australian poverty and policy since Henderson's landmark report. Panellists: Cassandra Goldie, Bruce Bradbury, Kay Cook, Jeremy Poxon www.eventbrite.com.au/e/unfinished...

Unfinished Business: 50 Years After Henderson

An initiative of the Australian Social Policy Association and the Australian Journal of Social Issues, as part of Anti-Poverty Week.

www.eventbrite.com.au

Stephen Nuñez @socio-steve.bsky.social · Sep 25

At some point people need to learn that the poverty rate is mostly about the safety net or lack thereof and not the economy. The economy was doing historically well, real wages rising for the first time in decades, etc. And poverty rose because pandemic safety policies expired.

Jacobin @jacobinmag.bsky.social · Sep 24

A new Census Bureau report establishes that poverty increased over the course of the Biden administration.

The data is yet another rebuke to the politicians and commentators who insisted economic conditions under Joe Biden were great.

It’s Official: Poverty Got Worse Under Joe Biden

A new Census Bureau report establishes that poverty increased over the course of the Biden administration. The data is yet another rebuke to the politicians and commentators who insisted economic conditions under Joe Biden were great.

jacobin.com

Aaron Sojourner @aaronsojourner.org · Sep 24

More evidence, on top of Baby's First Years experiment, that cash alone in early childhood does not improve children's early developmental trajectory.

In contrast, there is strong evidence that policies that improve children's care experiences do improve their early development & adult outcomes.

The Institute for Fiscal Studies @theifs.bsky.social · Sep 24

Even zooming in on poorer groups who are more likely to see larger income falls due to the policy, we still see no evidence of adverse effects on school readiness.

Chart shows the proportion of third and subsequent children achieving a good level of development by age 5, by date of birth relative to 6 April 2017, for families previously entitled to free school meals and those born in the 20% most deprived areas. Title states: "Children in families already entitled to free school meals or born in the poorest areas also saw no significant impact on school readiness from the two-child limit."

Bruce Bradbury @brucebradbury.bsky.social · Sep 2

And, of course: xkcd.com/552/

Correlation

xkcd.com

Bruce Bradbury @brucebradbury.bsky.social · Aug 31

"it is horribly bad news for the products of our education system. In order to be productive with AI, they need to obtain skills and experience that appear to come from learning by doing, except that they may not have the opportunities to ‘do.’ Second, wage dispersion is going to increase further"

Joshua Gans @joshgans.bsky.social · Aug 30

If AI and workers were strong complements, what would we see?, by @joshgans open.substack.com/pub/joshuaga... Does the latest data tell us AI is a substitute for human work? Nope. Is that comforting? Also nope. My comments on @erikbryn and co. (and @Noahpinion

If AI and workers were strong complements, what would we see?

The answer is pretty much what the initial data is showing

open.substack.com

Joshua Gans @joshgans.bsky.social · Aug 30

If AI and workers were strong complements, what would we see?, by @joshgans open.substack.com/pub/joshuaga... Does the latest data tell us AI is a substitute for human work? Nope. Is that comforting? Also nope. My comments on @erikbryn and co. (and @Noahpinion

If AI and workers were strong complements, what would we see?

The answer is pretty much what the initial data is showing

open.substack.com

Bruce Bradbury @brucebradbury.bsky.social · Aug 30

Crisis in Indonesia: "This is not yet about Prabowo. ... for the time being, this movement has been directed at the parliamentarians and the police."

Erin Cook @erincook.bsky.social · Aug 30

I have tried my very hardest to make this read about Indonesia's last 48 hours as accessible and easy to digest as possible. The goal here is that there is NO assumed knowledge (except for Jokowi, i think I always assume him sorry)
darimulut.substack.com/p/indonesia-...

🇮🇩 Indonesia shakes as protests turn to police violence and death

There's no coming back from this one

darimulut.substack.com

Erin Cook @erincook.bsky.social · Aug 30

I have tried my very hardest to make this read about Indonesia's last 48 hours as accessible and easy to digest as possible. The goal here is that there is NO assumed knowledge (except for Jokowi, i think I always assume him sorry)
darimulut.substack.com/p/indonesia-...

🇮🇩 Indonesia shakes as protests turn to police violence and death

There's no coming back from this one

darimulut.substack.com

Bruce Bradbury @brucebradbury.bsky.social · Aug 28

As an academic, I'm unsure whether my interests most align with those of the publishers of my papers, or with the LLMs. Making a small contribution to the corpus of human knowledge maintained by the LLMs doesn't seem such a bad outcome.

Bruce Bradbury @brucebradbury.bsky.social · Aug 27

No, Office automatically inserts en dashes. Typesetters are the ones who typically introduce the em dashes.

Bruce Bradbury @brucebradbury.bsky.social · Aug 21

Paul Krugman's substack series on Inequality is now all up on the Stone Center website:
stonecenter.gc.cuny.edu/tag/people-p...
(originally on his substack: paulkrugman.substack.com)

Paul Krugman Archives - Stone Center on Socio-Economic Inequality

stonecenter.gc.cuny.edu

Bruce Bradbury @brucebradbury.bsky.social · Aug 21

Exactly. There was tendency in policy research to think that longitudinal data would solve all our causal questions. I think (hope?) we are well past this now.

Julia M. Rohrer @dingdingpeng.the100.ci · Aug 21

It's like an association, but more causal.

This reasoning is very prevalent in psych as well (in particular when it comes to "lagged effects", aka lagged associations, and "within-person associations") which is why we wrote a paper about it:

journals.sagepub.com/doi/10.1177/...

These Are Not the Effects You Are Looking for: Causality and the Within-/Between-Persons Distinction in Longitudinal Data Analysis

In psychological science, researchers often pay particular attention to the distinction between within- and between-persons relationships in longitudinal data analysis. Here, we aim to clarify the relationship between the within- and between-persons distinction and causal inference and show that the distinction is informative but does not play a decisive role in causal inference. Our main points are threefold. First, within-persons data are not necessary for causal inference; for example, between-persons experiments can inform about (average) causal effects. Second, within-persons data are not sufficient for causal inference; for example, time-varying confounders can lead to spurious within-persons associations. Finally, despite not being sufficient, within-persons data can be tremendously helpful for causal inference. We provide pointers to help readers navigate the more technical literature on longitudinal models and conclude with a call for more conceptual clarity: Instead of letting statistical models dictate which substantive questions researchers ask, researchers should start with well-defined theoretical estimands, which in turn determine both study design and data analysis.

Julia M. Rohrer @dingdingpeng.the100.ci · Aug 21

It's like an association, but more causal.

This reasoning is very prevalent in psych as well (in particular when it comes to "lagged effects", aka lagged associations, and "within-person associations") which is why we wrote a paper about it:

journals.sagepub.com/doi/10.1177/...

Bruce Bradbury @brucebradbury.bsky.social · Aug 21

"the Australian government [should] establish an open, transparent and rigorous process to develop an official poverty measurement framework as soon as possible." Our paper in the latest issue of the Australian Economic Review onlinelibrary.wiley.com/toc/14678462...